Analysis of target detection performance for wireless sensor networks

In surveillance and tracking applications, wireless sensor nodes collectively monitor the existence of intruding targets. In this paper, we derive closed form results for predicting surveillance performance attributes, represented by detection probability and average detection delay of intruding targets, based on tunable system parameters, represented by node density and sleep duty cycle. The results apply to both stationary and mobile targets, and shed light on the fundamental connection between aspects of sensing quality and deployment choices. We demonstrate that our results are robust to realistic sensing models, which are proposed based on experimental measurements of passive infrared sensors. We also validate the correctness of our results through extensive simulations.